Robust strategies for hierarchical production planning

Hierarchical production planning is widely used in practice, the main reasons being that for mid-term planning less data is available than for short-term planning and that creating a detailed plan over a long time horizon is too time- consuming. Usually the different planning levels are treated separately; it is, however, evident that the dependencies between the planning levels are not negligible. Therefore, considering the dependencies and choosing the right methods could leverage a high potential for improving the planning strategies. In scientific literature numerous papers about planning methods for different levels can be found, many of which are used in practice. But as far as a systematic and extensive treatment of the interdependencies between the planning methods of different decision levels is concerned, little research has yet been done. For this project a situation typical for automotive suppliers will be used as a test case. The entire planning process from master planning down to the shop floor execution will be investigated, based on a stochastic environment (including market, machines, and workers), and taking into account simultaneously the two divergent objectives costs and service level. For that purpose various hierarchical planning strategies will be implemented and tested using a specially developed simulation environment. Special attention will be paid to the selection and optimization of the planning parameters. In order to obtain a better estimation of the planning strategies` potential, optimization models for a global planning problem will be developed and their solutions (exact ones for simplified scenarios, good approximations for more complex situations) will be compared to the hierarchical planning strategies with respect to their efficiency. Furthermore, analytical approaches will be developed in order to validate the empirical results, at least for simpler test instances, and to enhance the understanding of the dependencies between a strategy`s parameter values and performance indicators (e.g. costs and service level). The overall goal is to identify robust planning strategies for different production situations. This study will generate a set of rules for choosing suitable planning methods and parameters depending on the prevailing environmental conditions. Additionally, validity ranges for these choices can be used to determine when a change of planning strategy becomes necessary. This project will deliver an essential contribution to the improvement of the coordination of planning levels in a stochastic environment and represents an important step towards the application of the research results to real-world production planning problems. In the mid-term, the project may help to improve the planning stability of production planning systems used in practice.

To ensure long-term success on the market, companies should implement customer requirements economically. Proactive production planning and control is especially important to ensure that the right amount is cost-efficiently produced in the right quality at the right time. Hierarchical production planning systems are used to support these planning tasks. Depending on the maturity, long-, medium- and short-term levels are distinguished whereby the upper level defines the framework for the lower level. Long-term planning deals with sales planning and allocation of resources such equipment investments. For the medium-term planning, production orders are generated and staffing needs are defined. The main task for short-term planning is capacity setting and execution of production orders particularly focusing on dispatching. The project investigates the interrelations of the three planning levels. A simulation model has been developed applying the above-mentioned three-stage hierarchical production planning structure. Conducted simulation studies have identified a lack of capacity adjustment over the planning levels. Therefore, new methods have been developed on a long-, medium and short-term level and existing methods have been extended. The simulation model is the basis for the FFG follow up project Simgen generating efficient simulation models for a broad number of companies. In particular, an analytical model has been developed for long-term capacity investment decisions including customer required lead-time and necessary WIP. For the medium-term level, the traditional method of material requirement planning (MRP) has been extended by adding an integrated capacity restriction. A concept has been created for determining planned lead times, which are necessary for medium-term planning, depending on operational dispatching rules. Simulation-based optimization is used to identify good settings for planning parameters such as lot sizes or safety stocks. Results being published for the scientific community in peer-reviewed publications and have been presented at conferences. Furthermore, the results will be available as part of the follow-up FFG project Simgen for a broader audience as well as for manufacturing companies.